Analyzing the Impact of Climate Change With Major Emphasis on Pollution: A Comparative Study of ML and Statistical Models in Time Series Data

TL;DR

The paper addresses how pollution drives climate change and compares ML (LSTM, Spike Neural Networks) and statistical (ARIMA/SARIMA) time-series models using the Earth Surface Temperature Data and Global CO2 Dataset. It evaluates the ability of different models to capture CO2-emission–temperature relationships and to forecast future climate conditions up to $2100$. The study introduces temporal-encoding Spike Neural Networks, SARIMAX with exogenous CO2, and a city-level validation framework with 80/20 splits and multiple metrics, highlighting the strengths and limitations of each approach. Overall, the results show that traditional statistical models can be highly effective for climate-time-series forecasting, while ML methods offer complementary insights and can be enhanced to better handle pollution signals, with significant implications for climate policy and pollution mitigation.

Abstract

Industrial operations have grown exponentially over the last century, driving advancements in energy utilization through vehicles and machinery.This growth has significant environmental implications, necessitating the use of sophisticated technology to monitor and analyze climate data.The surge in industrial activities presents a complex challenge in forecasting its diverse environmental impacts, which vary greatly across different regions.Aim to understand these dynamics more deeply to predict and mitigate the environmental impacts of industrial activities.

Figures (8)

• Figure 1: Spiking Neural Network Architecture. Model architecture depicted for Spike Neural Network Architecture utilized to forecast global land temperatures.
• Figure 2: This figure represents the validation prediction of predicted versus test data for ARIMA
• Figure 3: Spiking Neural Network Training/Validation LossThe figure represents the training and validation loss over 50 epochs of the training model
• Figure 4: Spiking Neural Network Binary ForecastsThe figure represents the binary forecasts of the spike neural network model with a threshold of 0.5
• Figure 5: Validation Loss Curve for Jiddah ModelThe figure shows the validation loss curve on the validation data for the model trained on the Temperature data for Jaddih